The Gram stain is a common method for diagnosing bacterial vaginosis. However, Gram stain results are not routinely available right away, and treatment cannot be initiated at the time of the office visit. The Amstel criteria are used to quickly diagnose bacterial vaginosis. These criteria require at least three findings (thin homogenous vaginal discharge, a vaginal pH greater than 4.5, a positive “whiff” test, or a saline wet preparation that microscopically shows clue cells). Gutman and colleagues examined whether further simplifying the diagnosis by using fewer criteria would be as effective as using three criteria.
The authors studied samples from 269 women who received speculum examinations at a university clinic. Women with large amounts of vaginal bleeding at the time of vaginal examination were excluded from the study. Demographic and gynecologic data were collected for all participants. Examiners collected all four Amstel variables from the specimens using the Gram stain as the standard test.
Overall, 104 (38.7 percent) of participants’ Gram stains were positive for bacterial vaginosis. Women with bacterial vaginosis were an average of two years older than women who did not have vaginosis. Women with positive Gram stains were significantly more likely to be pregnant and to report vaginal discharge and odor compared with women who had negative Gram stains. Of the Amstel variables, vaginal pH greater than 4.5 had the highest sensitivity (89 percent), followed by thin homogenous discharge (79 percent), presence of clue cells (74 percent), and positive amine odor (67 percent). Conversely, positive amine odor had the highest specificity (93 percent), followed by presence of clue cells (86 percent), pH greater than 4.5 (74 percent), and thin homogenous discharge (54 percent).
The authors compared the conventional Amstel criteria with various other criteria cutoff points and combinations. A pH of 5.0 or greater had a sensitivity of 83 percent and a specificity of 82 percent for detecting bacterial vaginosis. Lowering the pH cutoff to 4.5 improved sensitivity to 89 percent and maintained a specificity of 74 percent. The best combination of sensitivity (74 percent) and specificity (86 percent) for presence of clue cells was obtained with more than 20 percent clue cells on a saline wet preparation. Combining any two findings improved sensitivity and specificity and was diagnostically equivalent to that of the conventional Amstel criteria.
The authors conclude that using any two of the four clinical findings is equivalent to the current recommendation of using at least three. They recommend that physicians initially perform a pH test, followed by any of the other tests. If the pH is 4.5 or greater and one other test is positive, the patient can be diagnosed with bacterial vaginosis and treated accordingly.